Automobiles equipped with speech-recognition and text-to-speech capabilities simplify tasks that would otherwise take away a driver's attention from driving. Current Voice Recognition (VR) systems allow a driver to tune to a broadcast station by using a VR command to tune to a broadcast frequency by speaking the broadcast frequency. As the VR system does not recognize broadcast station names, the current VR system is not very user friendly. Currently there are a lot of broadcast stations and Points-of-Interests (POIs) with common or similar names throughout the country. For example, the number of AM and FM stations in the United States is presently greater than 15,000 and the number is still growing. When searching for a broadcast station name or POI with a large dictionary two problems arise: high misrecognition rates due to large dictionary, and disambiguating a large number of matches to similar sounding broadcast stations and POIs, in particular, when determining a user's intent. As the number of stations grows, the problems of misrecognition and disambiguation get worse.